skip to main content

This content will become publicly available on June 18, 2024

Title: Multimodaltrace: Deepfake Detection using Audiovisual Representation Learning
By employing generative deep learning techniques, Deepfakes are created with the intent to create mistrust in society, manipulate public opinion and political decisions, and for other malicious purposes such as blackmail, scamming, and even cyberstalking. As realistic deepfake may involve manipulation of either audio or video or both, thus it is important to explore the possibility of detecting deepfakes through the inadequacy of generative algorithms to synchronize audio and visual modalities. Prevailing performant methods, either detect audio or video cues for deepfakes detection while few ensemble the results after predictions on both modalities without inspecting relationship between audio and video cues. Deepfake detection using joint audiovisual representation learning is not explored much. Therefore, this paper proposes a unified multimodal framework, Multimodaltrace, which extracts learned channels from audio and visual modalities, mixes them independently in IntrAmodality Mixer Layer (IAML), processes them jointly in IntErModality Mixer Layers (IEML) from where it is fed to multilabel classification head. Empirical results show the effectiveness of the proposed framework giving state-of-the-art accuracy of 92.9% on the FakeAVCeleb dataset. The cross-dataset evaluation of the proposed framework on World Leaders and Presidential Deepfake Detection Datasets gives an accuracy of 83.61% and 70% respectively. The study also provides insights into how the model focuses on different parts of audio and visual features through integrated gradient analysis  more » « less
Award ID(s):
Author(s) / Creator(s):
Date Published:
Journal Name:
IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Vancouver, CA, 2023
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Easy access to audio-visual content on social media, combined with the availability of modern tools such as Tensorflow or Keras, and open-source trained models, along with economical computing infrastructure, and the rapid evolution of deep-learning (DL) methods have heralded a new and frightening trend. Particularly, the advent of easily available and ready to use Generative Adversarial Networks (GANs), have made it possible to generate deepfakes media partially or completely fabricated with the intent to deceive to disseminate disinformation and revenge porn, to perpetrate financial frauds and other hoaxes, and to disrupt government functioning. Existing surveys have mainly focused on the detection of deepfake images and videos; this paper provides a comprehensive review and detailed analysis of existing tools and machine learning (ML) based approaches for deepfake generation, and the methodologies used to detect such manipulations in both audio and video. For each category of deepfake, we discuss information related to manipulation approaches, current public datasets, and key standards for the evaluation of the performance of deepfake detection techniques, along with their results. Additionally, we also discuss open challenges and enumerate future directions to guide researchers on issues which need to be considered in order to improve the domains of both deepfake generation and detection. This work is expected to assist readers in understanding how deepfakes are created and detected, along with their current limitations and where future research may lead. 
    more » « less
  2. A deepfake is content or material that is synthetically generated or manipulated using artificial intelligence (AI) methods, to be passed off as real and can include audio, video, image, and text synthesis. The key difference between manual editing and deepfakes is that deepfakes are AI generated or AI manipulated and closely resemble authentic artifacts. In some cases, deepfakes can be fabricated using AI-generated content in its entirety. Deepfakes have started to have a major impact on society with more generation mechanisms emerging everyday. This article makes a contribution in understanding the landscape of deepfakes, and their detection and generation methods. We evaluate various categories of deepfakes especially in audio. The purpose of this survey is to provide readers with a deeper understanding of (1) different deepfake categories; (2) how they could be created and detected; (3) more specifically, how audio deepfakes are created and detected in more detail, which is the main focus of this paper. We found that generative adversarial networks (GANs), convolutional neural networks (CNNs), and deep neural networks (DNNs) are common ways of creating and detecting deepfakes. In our evaluation of over 150 methods, we found that the majority of the focus is on video deepfakes, and, in particular, the generation of video deepfakes. We found that for text deepfakes, there are more generation methods but very few robust methods for detection, including fake news detection, which has become a controversial area of research because of the potential heavy overlaps with human generation of fake content. Our study reveals a clear need to research audio deepfakes and particularly detection of audio deepfakes. This survey has been conducted with a different perspective, compared to existing survey papers that mostly focus on just video and image deepfakes. This survey mainly focuses on audio deepfakes that are overlooked in most of the existing surveys. This article's most important contribution is to critically analyze and provide a unique source of audio deepfake research, mostly ranging from 2016 to 2021. To the best of our knowledge, this is the first survey focusing on audio deepfakes generation and detection in English. 
    more » « less
  3. Deepfakes, or synthetic audiovisual media developed with the intent to deceive, are growing increasingly prevalent. Existing methods, employed independently as images/patches or jointly as tubelets, have, up to this point, typically focused on spatial or spatiotemporal inconsistencies. However, the evolving nature of deepfakes demands a holistic approach. Inspection of a given multimedia sample with the intent to validate its authenticity, without adding significant computational overhead has, to date, not been fully explored in the literature. In addition, no work has been done on the impact of different inconsistency dimensions in a single framework. This paper tackles the deepfake detection problem holistically. HolisticDFD, a novel, transformer-based, deepfake detection method which is both lightweight and compact, intelligently combines embeddings from the spatial, temporal and spatiotemporal dimensions to separate deepfakes from bonafide videos. The proposed system achieves 0.926 AUC on the DFDC dataset using just 3% of the parameters used by state-ofthe-art detectors. An evaluation against other datasets shows the efficacy of the proposed framework, and an ablation study shows that the performance of the system gradually improves as embeddings with different data representations are combined. An implementation of the proposed model is available at: 
    more » « less
  4. Deepfakes represent the generation of synthetic/fake images or videos using deep neural networks. As the techniques used for the generation of deepfakes are improving, the threats including social media disinformation, defamation, impersonation, and fraud are becoming more prevalent. The existing deepfakes detection models, including those that use convolution neural networks, do not generalize well when subjected to multiple deepfakes generation techniques and cross-corpora setting. Therefore, there is a need for the development of effective and efficient deepfakes detection methods. To explicitly model part-whole hierarchical relationships by using groups of neurons to encode visual entities and learn the relationships between real and fake artifacts, we propose a novel deep learning model efficient-capsule network (E-Cap Net) for classifying the facial images generated through different deepfakes generative techniques. More specifically, we introduce a low-cost max-feature-map (MFM) activation function in each primary capsule of our proposed E-Cap Net. The use of MFM activation enables our E-Cap Net to become light and robust as it suppresses the low activation neurons in each primary capsule. Performance of our approach is evaluated on two standard, largescale and diverse datasets i.e., Diverse Fake Face Dataset (DFFD) and FaceForensics++ (FF++), and also on the World Leaders Dataset (WLRD). Moreover, we also performed a cross-corpora evaluation to show the generalizability of our method for reliable deepfakes detection. The AUC of 99.99% on DFFD, 99.52% on FF++, and 98.31% on WLRD datasets indicate the effectiveness of our method for detecting the manipulated facial images generated via different deepfakes techniques. 
    more » « less
  5. The use of audio and video modalities for Human Activity Recognition (HAR) is common, given the richness of the data and the availability of pre-trained ML models using a large corpus of labeled training data. However, audio and video sensors also lead to significant consumer privacy concerns. Researchers have thus explored alternate modalities that are less privacy-invasive such as mmWave doppler radars, IMUs, motion sensors. However, the key limitation of these approaches is that most of them do not readily generalize across environments and require significant in-situ training data. Recent work has proposed cross-modality transfer learning approaches to alleviate the lack of trained labeled data with some success. In this paper, we generalize this concept to create a novel system called VAX (Video/Audio to 'X'), where training labels acquired from existing Video/Audio ML models are used to train ML models for a wide range of 'X' privacy-sensitive sensors. Notably, in VAX, once the ML models for the privacy-sensitive sensors are trained, with little to no user involvement, the Audio/Video sensors can be removed altogether to protect the user's privacy better. We built and deployed VAX in ten participants' homes while they performed 17 common activities of daily living. Our evaluation results show that after training, VAX can use its onboard camera and microphone to detect approximately 15 out of 17 activities with an average accuracy of 90%. For these activities that can be detected using a camera and a microphone, VAX trains a per-home model for the privacy-preserving sensors. These models (average accuracy = 84%) require no in-situ user input. In addition, when VAX is augmented with just one labeled instance for the activities not detected by the VAX A/V pipeline (~2 out of 17), it can detect all 17 activities with an average accuracy of 84%. Our results show that VAX is significantly better than a baseline supervised-learning approach of using one labeled instance per activity in each home (average accuracy of 79%) since VAX reduces the user burden of providing activity labels by 8x (~2 labels vs. 17 labels).

    more » « less